CN108805957B - Vector diagram generation method and system based on bitmap image self-adaptive segmentation - Google Patents
Vector diagram generation method and system based on bitmap image self-adaptive segmentation Download PDFInfo
- Publication number
- CN108805957B CN108805957B CN201810578639.6A CN201810578639A CN108805957B CN 108805957 B CN108805957 B CN 108805957B CN 201810578639 A CN201810578639 A CN 201810578639A CN 108805957 B CN108805957 B CN 108805957B
- Authority
- CN
- China
- Prior art keywords
- region
- image
- edge
- area
- category
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
- G06T11/20—Drawing from basic elements, e.g. lines or circles
- G06T11/203—Drawing of straight lines or curves
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a vector diagram generation method and a vector diagram generation system based on bitmap image self-adaptive segmentation. And then, carrying out region segmentation processing on the bitmap image, merging the pixels with similar values into the same category according to the value of each pixel of the image, obtaining a region segmentation map of the image through multiple iterations, and merging the region with an excessively small area into a neighboring region. And obtaining a region mark map of the segmentation map, extracting the edge of each region by adopting an optimized edge model and an extraction algorithm, and fitting the edge line segment to determine an edge curve of each region. And finally, calculating the color average value of the source image corresponding to each category area in the area label image in the area, and taking the average value as the color of the current area to form a color list corresponding to the area. And generating a corresponding vector image file according to the edge curve data and the color list information of each region.
Description
Technical Field
The invention relates to the technical field of image processing and analysis, data processing and computer graphics, in particular to a vector diagram generation method and a vector diagram generation system based on bitmap image adaptive segmentation, which are used for quickly converting a bitmap into a vector diagram.
Background
Vector graphics, also known as object-oriented images, are represented using mathematical formula-based geometric image elements such as points, lines, rectangles, polygons, primitives, and arcs, which are referred to as objects. Each object is a self-contained entity having attributes of color, shape, outline, size, and position.
The vector diagram has the following advantages: the occupied storage space is small, and compared with a bitmap, the vector diagram only retains color information of lines and graphic blocks; independent of resolution, only the complexity of the image. The image is not distorted during zooming, rotating and deforming operations; high resolution printing is often used for designs, logos and text designs.
The current vector diagram generation methods mainly focus on the following two ways.
Firstly, the production is carried out by using commercial drawing software (such as Corel Draw, Adobe Illustrator and the like), personnel engaged in the production needs to be trained to acquire related use skills, meanwhile, the production process is long, and the effect of the generated vector image is often lower than the level of the production personnel.
And secondly, the software developer realizes the vectorization of a certain graph by programming by using the computer graphics principle. This generation method has the uniqueness of generating an image, and one implementation algorithm is usually for one or a class of vector graphics. With the number and kinds of vector graphics required to be designed, the energy and time consumed by developers will increase in multiples, and the development cost will be higher and higher.
As described above, a vector diagram generation method with fast operation and strong versatility is not currently formed. Therefore, it is desirable to provide a vector image generation method and system based on bitmap image adaptive segmentation.
Disclosure of Invention
The invention provides a vector diagram generation method based on bitmap image self-adaptive segmentation, which comprises steps S1-S3.
Step S1: and carrying out filtering operation on the bitmap image by adopting a bilateral filtering algorithm to remove random noise interference in the image.
Step S2: all pixel points of the bitmap image are initially set to be a category, the pixel points with similar values are combined into a category according to the characteristics of each pixel point, so that the image forms different regions, the regions with the too small area are combined according to the area of the regions to form a region segmentation image, and finally the pixel points in the different regions in the segmentation image are endowed with corresponding category numbers to obtain an initial region marking image.
Step S3: and extracting the coordinate data of the edge points of each region in the region mark image by adopting an optimized edge model and an extraction algorithm, and fitting the coordinate point data of the edge line segment by adopting a cubic Bezier curve to determine the edge curve of each region.
Step S4: and calculating the color average value of the source image corresponding to each category region in the region label image in the region, and taking the average value as the color of the current region.
Preferably, in step S1, the bilateral filtering algorithm adds the color difference weight on the basis of the gaussian filter, and retains the edge information of the image.
Preferably, step S2 further includes steps S21 to S23.
Step S21: and setting each pixel point of the image as a category, setting the initial intra-category difference of all the categories, traversing all the categories, and calculating the inter-category difference between the categories and the adjacent categories.
Step S22: traversing all adjacent classes, merging the two classes into one class when the difference between the two adjacent classes is smaller than the intra-class difference of the two classes, and repeating the traversing process until all the classes cannot be merged;
step S23: the area values of all classes, i.e. the number of pixels comprised by a class, are calculated. And traversing all classes, and merging the classes into the adjacent class with the minimum difference when the area value of a certain class is smaller than a preset value until the area values of all classes are larger than or equal to the preset value.
Step S24: and distinguishing each category area in the image by giving different category numbers so as to obtain an area marking image for vector diagram generation, wherein each marking area represents a graphic element of the vector diagram.
Preferably, in step S3, an optimized edge model and an optimized edge extraction algorithm are used, and when an edge is extracted, the upper edge point and the left edge point of the region are extracted in an inner edge tracking manner, and the lower edge point and the right edge point of the region are extracted in an outer edge tracking manner, so as to ensure that the common edges of the adjacent regions are consistent.
Preferably, the step S3 of extracting the coordinate data of the edge point of each region includes steps S311 to S312.
Step S311: acquiring an intersection point list of the region edge line and other region edge lines, starting from a first intersection point of the region, tracking the region edge until reaching a second intersection point of the region, which is a first edge line segment of the region;
step S312: and repeating the tracking process from the previous intersection point to the next intersection point in the previous step by taking the second intersection point as a starting point, and continuously acquiring the edge line segment of the area until the tracking ending condition is met. The tracking ending condition is that when the first intersection point of the area is tracked again, an edge line segment which takes the point as the starting point is obtained, if the line segment is the same as the first edge line segment, the tracking algorithm is ended, otherwise, the tracking process is continuously executed.
Preferably, the merged marking process is performed on the common edges of the neighboring regions in the data structure, so that only one fitting process needs to be performed on the common edges, and the processing process is accelerated.
Preferably, the step S3 of fitting the edge line segment with a cubic bezier curve to determine the edge curve of each region includes steps S321 to S323.
Step S321: setting a maximum threshold value of dispersion of each point of the fitting result and the input data, determining two end points of a curve, fitting the data points by using a least square method to obtain a middle control point, and obtaining a fitting result curve according to the middle control point;
step S322: and calculating the deviation value of each point between the fitting result and the input data to find the maximum deviation value. If the data points exceed the preset threshold, taking the data points corresponding to the current maximum separation value as breakpoints, dividing the input data into two sections of curves, and fitting the two sections of curves respectively;
step S323: and (3) repeating the operation of step (S322) on the fitting results of the two curves respectively, and enabling the fitting results to continuously approach the input data through continuous endpoint insertion and a piecewise fitting method until the deviation values of all points of the fitting result curve and the input data are within a preset threshold value, and finishing the fitting.
Preferably, step S4 further includes generating and saving a corresponding vector image file after respectively obtaining the edge curve line segment and the color value of each marked area.
The invention also provides a vector diagram generation system based on bitmap image self-adaptive segmentation, which comprises a noise elimination module, a segmentation module, a fitting module and a color endowing module, wherein the noise elimination module is connected with the segmentation module, the segmentation module is connected with the fitting module, the fitting module is connected with the color endowing module, wherein,
the de-noising module is used for carrying out filtering operation on the bitmap image and removing random noise interference in the image;
the segmentation module is used for setting each pixel point of the bitmap image into a category, combining the pixel points in sequence according to the characteristics of the pixel points to form different category regions, combining the categories with undersized areas according to the area of the category regions, and assigning category numbers to the pixel points of each category region in the bitmap image to obtain a region marker image;
the fitting module is used for extracting the edge of each region by adopting an optimized edge model and an extraction algorithm, fitting the edge line segment of each region by adopting a cubic Bezier curve and determining the edge curve of each region;
and the color-giving module is used for calculating the color average value of the source image corresponding to each category area in the area label image in the area, and taking the average value as the color value of the category area.
According to the vector diagram generation method and system based on the bitmap image self-adaptive segmentation, firstly, a bilateral filtering algorithm is adopted to carry out filtering operation on the bitmap image, and random noise interference in the image is removed. And then setting all pixel points of the bitmap image as a category, combining all the pixel points in the bitmap image into different regions in sequence, combining the regions according to the area of the regions, and giving category numbers to each pixel point in the bitmap image to obtain a region mark image. And then extracting the edge of each region by adopting an optimized edge model and an extraction algorithm, and fitting the edge line segment by adopting a cubic Bezier curve to determine the edge curve of each region. And finally, calculating the color average value of the source image corresponding to each category region in the region label image in the region, taking the average value as the color of the current region, and generating and storing a corresponding vector image file. Thus, automatic and fast conversion from bitmap to vector diagram is realized.
Other features and advantages of the present application will become more apparent from the detailed description of the embodiments of the present application when taken in conjunction with the accompanying drawings.
Drawings
FIG. 1a is a sample initial bitmap for conversion according to the preferred embodiment of the present invention;
FIG. 1b is a diagram of an image obtained by segmenting an initial bitmap using a vector graphics generation method based on bitmap image adaptive segmentation according to a preferred embodiment of the present invention;
FIG. 1c is a diagram of an image to be segmented and edge extracted using a vector graphics generation method based on bitmap image adaptive segmentation according to a preferred embodiment of the present invention;
FIG. 1d is a vector diagram converted using a vector diagram generation method based on bitmap image adaptive segmentation according to a preferred embodiment of the present invention;
FIG. 2a is an exemplary initial bitmap provided in accordance with a further preferred embodiment of the present invention;
FIG. 2b is the vector image corresponding to FIG. 2a transformed using a vector image generation method based on bitmap image adaptive segmentation;
FIG. 3a is an exemplary initial bitmap provided in accordance with a further preferred embodiment of the present invention;
FIG. 3b is the vector diagram corresponding to FIG. 3a converted using a vector diagram generation method based on bitmap image adaptive segmentation;
FIG. 4a is an exemplary initial bitmap provided in accordance with a further preferred embodiment of the present invention;
FIG. 4b is the vector image corresponding to FIG. 4a converted using the vector image generation method based on bitmap image adaptive segmentation;
FIG. 5a is an exemplary initial bitmap provided in accordance with a further preferred embodiment of the present invention;
FIG. 5b is the vector image corresponding to FIG. 5a converted using a vector image generation method based on bitmap image adaptive segmentation;
fig. 6 is a schematic structural diagram of a vector graphics generating system based on bitmap image adaptive segmentation according to a preferred embodiment of the present invention.
Detailed Description
The invention is further illustrated by the following figures and examples.
The invention provides a vector diagram generation method based on bitmap image self-adaptive segmentation, which comprises the steps of S1-S3.
Step S1: and carrying out filtering operation on the bitmap image by adopting a bilateral filtering algorithm to remove random noise interference in the image.
In particular, the operation of the step is convenient to be better applied to processing algorithms such as image clustering and image segmentation. The bilateral filtering algorithm is to add the color difference weight on the basis of the Gaussian filter, thereby better retaining the edge information of the image.
Step S2: initially setting each pixel point of the bitmap image as a category, merging the pixel points with similar values into a category according to the characteristics of each pixel point, so that the image forms regions of different categories, and merging the regions with undersized areas according to the area of the regions to form a region segmentation image. Finally, assigning pixel points of different areas in the segmented image to the category number of the area to obtain an initial area marking image;
in particular, the step relates to a graph theory segmentation algorithm, is a greedy clustering algorithm based on graph theory, and is commonly used for superpixel segmentation processing of images.
This step is broken down into S21 to S24.
Step S21: and initializing each pixel point in the image into a category, and setting the difference in the initial categories of all the categories. All classes are traversed and the inter-class differences with neighboring classes are calculated.
Specifically, the set of classes is referred to as a set of vertices V, and the initial intra-class difference Ci for each class (i.e., vertex) in V is set to 10. All classes (i.e., vertices) in V are traversed and their inter-class differences Dij are computed with neighboring classes (i.e., vertices). The intra-class difference is the root mean square value of the values of all points in the class, and the inter-class difference is the positive difference value of two adjacent classes.
Step S22: traversing all classes, merging the two classes into one class when the difference between the class and the adjacent class is smaller than the intra-class difference of the two classes, repeating the traversing process until all the classes cannot be merged, and finishing the classification;
with specific reference to the parameters, when Dij is less than both Ci and Cj, the two classes (i.e., vertices) are merged into one class.
Step S23: the area values of all classes, i.e. the number of pixels comprised by a class, are calculated. And traversing all the classes, merging the classes into the class with the minimum difference between the adjacent classes when the area value of a certain class is smaller than a preset threshold value until the area values of all the classes are larger than or equal to the preset value, and finishing the execution of the image region segmentation algorithm.
Specifically, the area maximum threshold is T, which determines the fineness of the generated vector diagram, and is generally set to be between 20 and 100 according to different fineness requirements.
Step S24: and distinguishing each category area in the image by giving different category numbers so as to obtain an area marking image for vector diagram generation, wherein each marking area represents a graphic element of the vector diagram.
Step S3: and extracting the edge of each region by adopting an optimized edge model and an extraction algorithm, and fitting the edge line segment by adopting a cubic Bezier curve to determine the edge curve of each region.
Specifically, edge information of each category area in the marked image is extracted by adopting an optimized edge model and an extraction algorithm, and a cubic Bezier curve is adopted for fitting edge line segments to obtain an edge curve of each category area, namely edge information data of each graphic element of the vector diagram.
As described above, the segmentation algorithm is a bitmap segmentation algorithm, and each region in the extracted region marker image is composed of pixels, so that the region edge is irregular. The shape of the region needs to be represented by a curve fitting algorithm using mathematically closed curve segments, i.e. fitting a number of cubic bezier curves to the edge lines of each primitive.
The traditional edge extraction algorithm extracts the outer edge or the inner edge of a certain area, but the common edge of adjacent areas generates different edge point data in the extraction process no matter the inner edge or the outer edge, and the common edge line segments of the adjacent areas generate different fitting results, so that an overlapping area or a blank area is generated between the adjacent areas. In the embodiment, an optimized edge model and an optimized edge extraction algorithm are adopted, when the edge is extracted, the upper edge point and the left edge point of the region are extracted in an inner edge mode, and the lower edge point and the right edge point of the region are extracted in an outer edge mode so as to be overlapped with the edge of the adjacent region, so that the consistency of the common edge of the adjacent regions is ensured.
The specific process is as follows:
step S311: acquiring an intersection point list of each region edge line and other region edge lines, starting from a first intersection point of the region, tracking the region edge until reaching a second intersection point of the region, which is a first edge line segment of the region;
step S312: and repeating the tracking process from the previous intersection point to the next intersection point in the previous step by taking the second intersection point as a starting point, and continuously acquiring the edge line segment of the area until the tracking ending condition is met. And the tracking ending condition is that when the first intersection point of the area is tracked again, an edge line segment taking the point as the starting point is obtained, if the line segment is the same as the first edge line segment, the tracking algorithm is ended, otherwise, the tracking process is continuously executed.
In step S312, in order to reduce the fitting time, the merging labeling process is performed on the common edges of the neighboring regions in the data structure, so that the fitting process only needs to be performed once on the common edges, thereby speeding up the processing procedure.
Cubic or quartic bezier curves are typically used in the edge curve segment fitting algorithm. The formula of the m-grade Bezier curve is as follows:
(1) in the formula: t is tiThe values of (A) are often discrete, tiCorresponds to a point q (t) on the curvei) Therefore, the number of values directly influences the fineness of the curve; m is the number of bezier curve steps; pkIs the kth control of the curveAnd (5) point making.
Here we use a cubic bezier curve:
q(ti)＝P0(1-ti)3+3P1(1-ti)2ti+3P2(1-ti)ti 2+P3ti 3(2) formula (II)
(2) In the formula: p0、P3Determining the head and tail positions of the Bezier curve as the end points (head and tail control points) of the Bezier curve; p1、P2For the control points of the bezier curve, the trend of the curve is determined.
The main objective of cubic bezier curve fitting is to find the control point P of the bezier curve1、P2Head and tail end point P0、P3Can be taken directly from the input data points. Finding curve control point P by least square method1、P2The difference between the fit and the input data points is minimized.
The least square method formula is as follows:
in the formula: s is the sum of the squares of the errors, piAre input data points.
Substituting equation (2) into equation (3):
s to P1、P2Derivative, with a partial derivative of 0:
solving equation (5) to obtain P1、P2The value of (c):
P1＝(A2C1-A12C2)/(A1A2-A12A12) (6) formula
P2＝(A1C2-A12C1)/(A1A2-A12A12) (7) formula
In the formula:
the fitting procedure is described below.
Step S323:
preferably, step S4 further includes generating and saving a corresponding vector image file after respectively obtaining the edge curve line segment and the color value of each marked area.
Step S321: setting a maximum threshold value of dispersion of each point of the fitting result and the input data, determining two end points of a curve, fitting the data points by using a least square method to obtain a middle control point, and obtaining a fitting result curve according to the middle control point;
specifically, a maximum threshold of dispersion (similarity threshold) of the fitting algorithm, that is, a maximum threshold of a distance between a fitting result curve and an input data curve is determined. And determines two end points, i.e. the first point P of the input points0And the last point P3. Fitting the data points by using a least square method to obtain a control point P1And P2The value of (c).
Step S322: and calculating the deviation value of each point between the fitting result and the input data to find the maximum deviation value. And if the data points exceed the preset threshold, taking the data points corresponding to the current maximum separation value as breakpoints, dividing the input data into two sections of curves, and then respectively performing fitting processing on the two sections of curves.
Step S323: and (3) repeating the operation of step (S322) on the fitting results of the two curves respectively, and enabling the fitting results to continuously approach the input data by continuously inserting end points and performing piecewise fitting until the deviation values of all the points of the fitting result curve and the input data are within a preset threshold value, and finishing the fitting.
Step S4: and calculating the color average value of the source image corresponding to each category region in the region mark image in the region, and taking the average value as the color of the current region.
In step S3, the edge information of each region in the region-divided image, i.e., the boundary curves corresponding to all the primitives in the vector diagram, is obtained. Next, the color feature of each region, that is, the color value corresponding to each primitive, needs to be obtained. And calculating the color average value of the source image corresponding to each category area in the area marked image by using the area marked image generated in the step, and taking the average value as the color of the current area so as to determine the color value of the corresponding primitive.
And after the boundary curve and the color value of each primitive are respectively obtained, generating a corresponding vector image file. Specifically, a vector image file corresponding to the bitmap can be generated according to the format specification requirement of the SVG, and the vector image file is stored in the format of the SVG.
According to the vector diagram generation method based on bitmap image adaptive segmentation provided by the preferred embodiment of the invention, the inventor makes a partial sample diagram conversion example.
As shown in fig. 1(a) to 1(d), the process of generating a vector diagram from preprocessing, image segmentation, edge fitting to primitive coloring by an initial bitmap is shown.
Next, test data is also recorded in table 1 with bitmap conversion of different pixels. The data shows that the higher the image pixels, the longer the converted picture takes, and the higher the resultant vector image sharpness.
| Initial bitmap | Generating vector diagrams | Image size (pixel) | Number of divided regions | Elapsed time (milliseconds) |
| FIG. 2(a) | FIG. 2(b) | 369*470 | 2413 | 13796ms |
| FIG. 3(a) | FIG. 3(b) | 560*408 | 1808 | 12132ms |
| FIG. 4(a) | FIG. 4(b) | 682*682 | 2081 | 11301ms |
| FIG. 5(a) | FIG. 5(b) | 1024*768 | 4805 | 46787ms |
TABLE 1
Fig. 6 is a schematic structural diagram of a vector graphics generating system based on bitmap image adaptive segmentation according to a preferred embodiment of the present invention. As shown in fig. 6, the vector diagram generating system based on bitmap image adaptive segmentation provided in this embodiment includes a noise elimination module 101, a segmentation module 102, a fitting module 103, and a color-adding module 104, where the noise elimination module 101 is connected to the segmentation module 102, the segmentation module 102 is connected to the fitting module 103, and the fitting module 103 is connected to the color-adding module 104. The denoising module 101 is configured to perform a filtering operation on the bitmap image to remove random noise interference in the image. The segmentation module 102 is configured to set each pixel point of the bitmap image as a category, sequentially merge the pixel points according to characteristics of the pixel points to form different regions, merge regions with an excessively small area according to an area of the region, and assign a category number to the pixel point of each region in the bitmap image to obtain a region marker image. The fitting module 103 is configured to extract edges of each region by using the optimized edge model and the extraction algorithm, and fit the edge line segments by using a cubic bezier curve to determine an edge curve of each region. The color-giving module 104 is configured to calculate an average value of colors of the source image in each category region in the region labeled image, and use the average value as the color of the category region.
In summary, according to the vector diagram generation method and system based on bitmap image adaptive segmentation provided by the preferred embodiment of the present invention, the method first performs a filtering operation on the bitmap image by using a bilateral filtering algorithm, so as to remove random noise interference in the image. And then, setting each pixel point of the bitmap image into a category, combining the pixel points in sequence according to the characteristics of the pixel points to form different category regions, combining the categories with undersized areas according to the area of the category regions, and assigning category numbers to the pixel points of each category region in the bitmap image to obtain a region marking image. And then extracting the edge of each region by adopting an optimized edge model and an extraction algorithm, and fitting the edge line segment by adopting a cubic Bezier curve to determine the edge curve of each region. And finally, calculating the color average value of the source image corresponding to each category area in the area label image in the area, taking the average value as the color of the current area, and generating a corresponding vector image file. Therefore, the vector graphics can be quickly generated only by inputting the bitmap image to be converted without manual operation.
The foregoing is illustrative of the preferred embodiments of the present invention and it will be appreciated by those skilled in the art that various modifications and adaptations can be made without departing from the principles of the invention and should be considered as within the scope of the invention.
Claims (8)
1. A vector graphics generation method based on bitmap image adaptive segmentation is characterized by comprising the following steps:
s1, filtering the bitmap image by adopting a bilateral filtering algorithm to remove random noise interference in the image;
s2, setting each pixel point of the bitmap image into a category, combining the pixel points with similar values into a category according to the characteristics of each pixel point, so that the image forms areas of different categories, combining the category areas with undersized areas to form a region segmentation image, and finally endowing the pixel points of different regions in the segmentation image with category numbers of the region to obtain a region marking image;
s3, extracting the edges of each region by adopting an optimized edge model and an extraction algorithm, and fitting the edge line segments by adopting a cubic Bezier curve to determine the edge curve of each region;
in step S3, fitting the edge line segment with a cubic bezier curve to determine an edge curve of each region, including:
s321, setting a maximum deviation threshold value of the fitting result and each point of the input data, determining two end points of a curve, fitting the data points by using a least square method to obtain a middle control point, and obtaining a fitting result curve according to the middle control point;
s322, calculating the dispersion value of each point between the fitting result and the input data, finding the maximum dispersion value, if the maximum dispersion value exceeds a preset threshold value, taking the data point corresponding to the current maximum dispersion value as a breakpoint, dividing the input data into two sections of curves, and fitting the two sections of curves respectively;
s323, repeating the operation of the step S322 on the fitting result of the two sections of curves respectively, and enabling the fitting result to continuously approach the input data through continuous insertion of endpoints and a sectional fitting method until the deviation values of all points of the fitting result curve and the input data are within a preset threshold value, and finishing the fitting;
and S4, calculating the color average value of the source image corresponding to each category number in the area label image in the area, and taking the average value as the color of the current area.
2. The vector graphics generating method based on bitmap image adaptive segmentation as claimed in claim 1, wherein in step S1, the bilateral filtering algorithm is to add color difference weight based on gaussian filter, and to retain the edge information in the image.
3. The vector graphics generating method based on bitmap image adaptive segmentation according to claim 1, wherein step S2 further comprises:
s21, setting each pixel point of the image as a category, setting the initial intra-category difference of all categories, traversing all the categories, and calculating the inter-category difference between the categories and the adjacent categories;
s22, traversing all adjacent classes, merging the two classes into one class when the difference between the two adjacent classes is smaller than the intra-class difference of the two classes, and repeating the traversing process until all the classes cannot be merged;
s23, calculating the area values of all classes, namely the number of pixels included in each class, traversing all the classes, and merging the classes into the class with the minimum difference between the adjacent classes when the area value of a certain class is smaller than a preset value until the area values of all the classes are larger than or equal to the preset value;
step S24: and distinguishing each category area in the image by giving different category numbers so as to obtain an area marking image for vector diagram generation, wherein each marking area represents a graphic element of the vector diagram.
4. The vector graphics generating method based on bitmap image adaptive segmentation as claimed in claim 1, wherein in step S3, using an optimized edge model and extraction algorithm, when extracting the edge, the upper edge point and the left edge point of the region are extracted in a manner of inner edge tracking, and the lower edge point and the right edge point of the region are extracted in a manner of outer edge tracking, so as to ensure that the common edges of adjacent regions are consistent.
5. The vector graphics generation method based on bitmap image adaptive segmentation of claim 1, wherein in step S3, extracting region edges comprises:
s311, acquiring an intersection point list of each region edge line and other region edge lines, starting from a first intersection point of the region, tracking the region edge until reaching a second intersection point of the region, wherein the second intersection point is a first edge line segment of the region;
and S312, repeating the tracking process from the previous intersection point to the next intersection point in the previous step by taking the second intersection point as a starting point, and continuously acquiring the edge line segment of the area until a tracking ending condition is met, wherein the tracking ending condition is that when the first intersection point of the area is tracked again, the edge line segment taking the point as the starting point is acquired, if the line segment is the same as the first edge line segment, the tracking algorithm is ended, otherwise, the tracking process is continuously executed.
6. The vector graphics generating method based on bitmap image adaptive segmentation as claimed in claim 5, wherein the merging labeling process is performed on the common edges of the neighboring regions in the data structure, so that only one fitting process is required for the common edges, and the processing process is accelerated.
7. The vector graphics generating method based on bitmap image adaptive segmentation as claimed in claim 1, wherein step S4 further comprises generating a corresponding vector graphics file after obtaining the edge curve line segment and the color value of each marked region respectively.
8. A vector diagram generating system based on bitmap image self-adaptive segmentation is characterized by comprising a noise elimination module, a segmentation module, a fitting module and a coloring module, wherein the noise elimination module is connected with the segmentation module, the segmentation module is connected with the fitting module, the fitting module is connected with the coloring module, the noise elimination module is connected with the segmentation module,
the de-noising module is used for carrying out filtering operation on the bitmap image and removing random noise interference in the image;
the segmentation module is used for setting each pixel point of the bitmap image into a category, combining the pixel points in sequence according to the characteristics of the pixel points to form different category regions, combining the categories with undersized areas according to the area of the category regions, and assigning category numbers to the pixel points of each category region in the bitmap image to obtain a region marker image;
the fitting module is used for extracting the edges of the regions by adopting an optimized edge model and an extraction algorithm, and fitting the edge line segments of the regions by adopting a cubic Bezier curve to determine the edge curve of each region;
and the color-giving module is used for calculating the color average value of the source image corresponding to each category area in the area label image in the area, and taking the average value as the color value of the category area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810578639.6A CN108805957B (en) | 2018-06-07 | 2018-06-07 | Vector diagram generation method and system based on bitmap image self-adaptive segmentation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810578639.6A CN108805957B (en) | 2018-06-07 | 2018-06-07 | Vector diagram generation method and system based on bitmap image self-adaptive segmentation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108805957A CN108805957A (en) | 2018-11-13 |
| CN108805957B true CN108805957B (en) | 2022-06-24 |
Family
ID=64087574
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810578639.6A Active CN108805957B (en) | 2018-06-07 | 2018-06-07 | Vector diagram generation method and system based on bitmap image self-adaptive segmentation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108805957B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110021054B (en) * | 2019-04-12 | 2023-07-28 | 青岛九维华盾科技研究院有限公司 | Plaque coloring method for speckle camouflage pattern design |
| CN110120080B (en) * | 2019-04-12 | 2024-01-05 | 青岛九维华盾科技研究院有限公司 | Method for rapidly generating main color of standard camouflage color |
| CN111127592B (en) * | 2019-12-25 | 2023-11-28 | 超级魔方(北京)科技有限公司 | Picture color filling method and device, electronic equipment and readable storage medium |
| CN111179370B (en) * | 2019-12-30 | 2023-06-20 | 北京猎豹网络科技有限公司 | Picture generation method and device, electronic equipment and storage medium |
| CN111210485B (en) | 2020-01-06 | 2023-03-28 | 北京字节跳动网络技术有限公司 | Image processing method and device, readable medium and electronic equipment |
| CN111242859B (en) * | 2020-01-07 | 2023-09-08 | 上海菲戈恩微电子科技有限公司 | Improved image processing method |
| CN111353957A (en) * | 2020-02-28 | 2020-06-30 | 北京三快在线科技有限公司 | Image processing method, image processing device, storage medium and electronic equipment |
| CN111462136B (en) * | 2020-03-31 | 2023-05-16 | 上海大学 | An Image Vectorization System and Method Based on Boundary Segmentation Algorithm |
| CN113744114B (en) * | 2020-05-27 | 2024-11-12 | 京东方科技集团股份有限公司 | Vector graphics drawing method, device and storage medium based on 8k video system |
| CN112181923B (en) * | 2020-09-23 | 2024-06-14 | 广东康派环创科技有限公司 | Bitmap acquisition method, bitmap acquisition device, computer equipment and storage medium |
| CN112215272A (en) * | 2020-09-29 | 2021-01-12 | 重庆大学 | Bezier curve-based image classification neural network attack method |
| CN113706648B (en) * | 2021-08-04 | 2024-03-19 | 武汉理工大学 | Thangka element vectorization method based on diffusion curve |
| CN114022483B (en) * | 2022-01-08 | 2022-03-25 | 南通欣斯特机械制造有限公司 | Injection molding flash area identification method based on edge characteristics |
| CN114820873B (en) * | 2022-05-13 | 2023-04-07 | 南京大学 | Sketch regularization method, device and medium based on painting |
| CN114661930B (en) * | 2022-05-24 | 2022-08-26 | 深圳顶匠科技有限公司 | Bitmap processing method, device and equipment based on extensible markup language |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007241878A (en) * | 2006-03-10 | 2007-09-20 | Toshiba Corp | Drawing apparatus, drawing method, and drawing program |
| CN103927727A (en) * | 2014-04-22 | 2014-07-16 | 成都品果科技有限公司 | Method for converting scalar image into vector image |
| CN104268330A (en) * | 2014-09-19 | 2015-01-07 | 华南理工大学 | A Bezier Fitting Method for Complicated Curves |
| CN105809699A (en) * | 2016-03-18 | 2016-07-27 | 中山大学 | Image segmentation based car window extraction method and system |
| CN107273608A (en) * | 2017-06-13 | 2017-10-20 | 中国石油大学(华东) | A kind of reservoir geology profile vectorization method |
| CN107704801A (en) * | 2017-08-18 | 2018-02-16 | 电子科技大学 | Curved lanes line detecting method based on segmented linear bonus point section Bezier |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1380543A (en) * | 2001-04-12 | 2002-11-20 | 清华大学 | Image segmentation and identification method in industrial radiation imaging system |
| US7212205B2 (en) * | 2002-11-12 | 2007-05-01 | Matsushita Electric Industrial Co., Ltd. | Curved surface image processing apparatus and curved surface image processing method |
-
2018
- 2018-06-07 CN CN201810578639.6A patent/CN108805957B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007241878A (en) * | 2006-03-10 | 2007-09-20 | Toshiba Corp | Drawing apparatus, drawing method, and drawing program |
| CN103927727A (en) * | 2014-04-22 | 2014-07-16 | 成都品果科技有限公司 | Method for converting scalar image into vector image |
| CN104268330A (en) * | 2014-09-19 | 2015-01-07 | 华南理工大学 | A Bezier Fitting Method for Complicated Curves |
| CN105809699A (en) * | 2016-03-18 | 2016-07-27 | 中山大学 | Image segmentation based car window extraction method and system |
| CN107273608A (en) * | 2017-06-13 | 2017-10-20 | 中国石油大学(华东) | A kind of reservoir geology profile vectorization method |
| CN107704801A (en) * | 2017-08-18 | 2018-02-16 | 电子科技大学 | Curved lanes line detecting method based on segmented linear bonus point section Bezier |
Non-Patent Citations (1)
| Title |
|---|
| 基于最小二乘法的分段三次曲线拟合方法研究;蔡山;《科学技术与工程》;20070228;第7卷(第3期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108805957A (en) | 2018-11-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108805957B (en) | Vector diagram generation method and system based on bitmap image self-adaptive segmentation | |
| Jiang et al. | Scfont: Structure-guided chinese font generation via deep stacked networks | |
| CN106384094B (en) | An automatic generation method of Chinese character library based on writing style modeling | |
| US9607391B2 (en) | Image object segmentation using examples | |
| CN104063723B (en) | The stroke restoring method and device of the Off-line Handwritten Chinese | |
| CN105931295B (en) | A method for extracting thematic information from geological maps | |
| CN111310760B (en) | Oracle Bone Inscription Text Detection Method Combining Local Prior Features and Deep Convolution Features | |
| CN102968764B (en) | Chinese character image inpainting method based on strokes | |
| CN111126127B (en) | High-resolution remote sensing image classification method guided by multi-level spatial context characteristics | |
| US8670615B2 (en) | Refinement of segmentation markup | |
| CN104537676A (en) | Gradual image segmentation method based on online learning | |
| CN112784531B (en) | Chinese character form and character library generation method based on deep learning and component splicing | |
| CN113158977B (en) | Image character editing method for improving FANnet generation network | |
| CN104834890B (en) | A kind of extracting method to word expression information in calligraphy work | |
| Yao et al. | Manga vectorization and manipulation with procedural simple screentone | |
| CN110889374A (en) | Seal image processing method and device, computer and storage medium | |
| CN109409211B (en) | Processing method, processing device and storage medium for Chinese character skeleton stroke segments | |
| CN105335685A (en) | Image identification method and apparatus | |
| CN112347288B (en) | Vectorization method of word graph | |
| CN113033656A (en) | Interactive hole exploration data expansion method based on generation countermeasure network | |
| CN111382745A (en) | Nail image segmentation method, device, equipment and storage medium | |
| CN109325483B (en) | Method and device for processing internal short pen section | |
| Manushree et al. | XCI-sketch: Extraction of color information from images for generation of colored outlines and sketches | |
| CN111476243A (en) | Image character recognition method and device | |
| Tekeli et al. | Shape and data driven texture segmentation using local binary patterns |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |